Canfield R L, Smith E G, Brezsnyak M P, Snow K L
Department of Human Development, Cornell University, USA.
Monogr Soc Res Child Dev. 1997;62(2):1-145.
This Monograph uses a developmental function approach to describe age-related change and individual differences in infant information processing during the first year of life. The Visual Expectation Paradigm (VExP) is used to measure speed of information processing, response variability, and expectancy formation. Eye-movement reaction times and anticipatory saccades were gathered from 13 infants assessed monthly from 2 to 9 months and then again at 12 months. Analysis of response patterns demonstrated the applicability of the paradigm throughout the age range studied. Converging operations strongly indicate that the traditional estimate of the minimum time required for infants to initiate a saccade to a peripheral stimulus may be as much as 100 milliseconds (ms) too long. Moreover, the newly estimated minimum of 133 ms does not appear to change during the 2-12-month period. Reanalysis of the present data and past research reveals that the new, shorter minimum reaction time is unlikely to affect findings based on mean reaction time. However, using the traditional minimum reaction time will inflate estimates of percentage anticipation, especially in infants older than 5 months. Group and individual growth curves are described through quantitative models of four variables: reaction time, standard deviation of reaction time, percentage anticipation, and anticipation latency. Developmental change in reaction time was best described by an asymptotic exponential function, and evidence for a local asymptote during infancy is presented. Variability in reaction time was found to decline with age, independent of mean reaction time, and was best described by a polynomial function with linear and quadratic terms. Anticipation showed little lawful change during any portion of the age span, but latency to anticipate declined linearly throughout the first year. Stability of individual differences was strong between consecutive assessments of mean reaction time. For nonconsecutive assessments, stability was found only for the 6-12-month period. Month-to-month stability was inconsistent for reaction-time variability and weak for both anticipation measures. Analyses of individual differences in growth curves were carried out using random regressions for the polynomial models. The only significant individual difference (in growth curves) was found for reaction-time variability. Parameter estimates from the exponential models for reaction time suggested two or three developmental patterns with different exponential trajectories. This finding indicates that the strong form of the exponential growth hypothesis, which states that processing speed develops at the same rate for all individuals, does not hold for the first year of life. In the concluding chapter, Grice's Variable Criterion Model (Grice, 1968) is used to integrate three key findings: regular age changes in mean reaction time and variability but no age change in the minimum reaction time. It is argued that the rate of growth of sensory-detection information is developmentally constant during much of the first year but that age changes occur in the level and spread of the distribution of response threshold values. The unique strengths of the paradigm are discussed, and future directions are suggested for further developing the paradigm itself and for using it as a tool to study broad issues in infant cognition.
本专题采用发展功能方法来描述一岁以内婴儿信息处理过程中与年龄相关的变化及个体差异。视觉期望范式（VExP）用于测量信息处理速度、反应变异性和期望形成。对13名婴儿进行了眼动反应时间和预期扫视测量，这些婴儿从2个月至9个月每月评估一次，然后在12个月时再次评估。对反应模式的分析表明该范式在所研究的整个年龄范围内都适用。多项研究结果有力地表明，传统上估计婴儿对外周刺激发起扫视所需的最短时间可能长了多达100毫秒（ms）。此外，新估计的133 ms的最短时间在2至12个月期间似乎没有变化。对现有数据和以往研究的重新分析表明，新的、更短的最短反应时间不太可能影响基于平均反应时间的研究结果。然而，使用传统的最短反应时间会夸大预期百分比的估计值，尤其是在5个月以上的婴儿中。通过四个变量的定量模型来描述组和个体的生长曲线：反应时间、反应时间的标准差、预期百分比和预期潜伏期。反应时间的发育变化最好用渐近指数函数来描述，并给出了婴儿期存在局部渐近线的证据。发现反应时间的变异性随年龄下降，与平均反应时间无关，最好用包含线性和二次项的多项式函数来描述。在整个年龄跨度的任何阶段，预期变化都很少有规律，但预期潜伏期在第一年中呈线性下降。连续评估平均反应时间时，个体差异的稳定性很强。对于非连续评估，仅在6至12个月期间发现了稳定性。反应时间变异性的逐月稳定性不一致，两种预期测量的稳定性都较弱。使用多项式模型的随机回归对生长曲线中的个体差异进行了分析。唯一显著的个体差异（在生长曲线中）是反应时间变异性。反应时间指数模型的参数估计表明存在两到三种具有不同指数轨迹的发育模式。这一发现表明，指数增长假说的强形式，即认为所有个体的处理速度以相同速度发展，在生命的第一年并不成立。在最后一章中，使用了格赖斯可变标准模型（格赖斯，1968）来整合三个关键发现：平均反应时间和变异性随年龄有规律变化，但最短反应时间没有年龄变化。有人认为，在第一年的大部分时间里，感觉检测信息的增长速度在发育上是恒定的，但反应阈值分布的水平和范围会随年龄变化。讨论了该范式的独特优势，并为进一步发展该范式本身以及将其用作研究婴儿认知广泛问题的工具提出了未来方向。
